'The world is not prepared:' How AI energy thirst might tap into geothermal power

Tech companies champing at the bit to create the latest, greatest generative AI models face an uncomfortable dilemma. Data-hungry models like ChatGPT and Google's Gemini rely on troves of digital material in data centers that require massive amounts of energy for processing and constant cooling. Some estimates suggest this swelling energy demand could account for nine percent of all US electricity by the end of the decade. That's up from around four percent today, a steep increase experts attribute in part to Big Tech's brewing generative AI arms race. Renewable energy sources like wind and solar aren't ready to meet demand alone. Instead, much of the new energy could come from fossil fuel sources which could undermine many of these companies' ambitious "net zero" and carbon neutral pledges.